whythismatters

>haven't heard of any Chinese models "escaping"

There was this incident that seemingly flew under the radar (52 days ago): https://news.ycombinator.com/item?id=49216185

My first impression is that things have improved. I encountered some "honesty" though :)

Great post, thank you! So, let's resolve ancient Near East chronology and determine an absolute date for the sack of Babylon?

I recently tried to revive my account after not having accessed it for a decade or so. They asked for my gov issued ID, to be processed by the Persona service. Um, hell no.

The juvenile nonchalance with which some Anthropic employees seem to be talking to their AI (wacky, sick, cook, ...) is truly bizarre.

The author mentions Angular in the comments as a personal blind spot

discussion of oai announcement: https://news.ycombinator.com/item?id=49790389

Don't worry, in its final reply the agent will always present one or two things "worth flagging" to keep you engaged

As some commenter in another thread pointed out OP is on a crusade, posting variations on this theme (and getting flagged) all the time: https://news.ycombinator.com/submitted?id=Bitu79

See also (hypothetical) Proto-Euphratean, "the lost language hidden within Sumerian": https://en.wikipedia.org/wiki/Proto-Euphratean_language

More on the upcoming priesthood and monasteries: https://news.ycombinator.com/item?id=49743400

Funny! Not sure if it's a cultural thing, but I experienced this phenomenon with Indian and Portuguese guys, and while I'm sure they're just trying to be polite, it's annoying as hell. Just ask the damn question!

I'm not affiliated, and did not actually try it, but your comment reminded me of the juggler coding agent, discussed here: https://news.ycombinator.com/item?id=48883305

>Each conversation is a Yjs CRDT tree. It can branch into sub-threads (recursively), and you can drill down, backtrack, edit, undo/redo, and inspect everything: tool calls, approvals, and the raw context JSON going to the model, etc.

>Submitted on 5 Feb 2026

worth flagging

>What comes next is a new kind of product we’re developing: opt-in experiences built around individual artists, where artists can choose to participate and get paid when they do, and fans get new ways to connect with the artists they love.

Actual title of the article: "Muse Give Up Their Social Media Handle to AI in Ironic Twist"

>slimy

I can't put my finger on it, but there's something off about this article, e.g. glossing over the opportunism (acting on "rumors"), drive-by claim about "strict safeguards [...] including monitoring and isolation", high horse attitude (we gave the guy a chance, we don't care about 1M USD, and while you fools are complaining we just tick this box and continue the pursuit of our noble goals for the benefit of humanity). I don't like it.

>a cached version of the internet

Interesting detail. A heavily pruned version, I assume?

The authors frame it as "a twisted and sadistic form of climate-change adaptation."

I read it, perhaps ironically, as colloquial/slang US English, like "lemme git to pushin' y'all!"

>this particular "persistence-model" was encrypted and locked away, even from OAI staff

source?

Does it mean they made 100 billion in profits? Cf. the AGI deal with Microsoft (https://news.ycombinator.com/item?id=47921248)

I successfully used it to re-slop Claude slop from 1+ year ago. Really puts things into perspective.

I tried the self hosted app some months ago, and while the vaults concept was nice for isolating e.g work and personal, some features were a bit clunky, something wrong with tagging or connecting contacts with notes iirc. In my opinion, it also had irrelevant features, I think it was mood tracking? Based on the experience, and the fact that the announced rewrite will be more "playful," I doubt I will try it again.

>planned to be temporary

Same for the Atomium in Brussels, Belgium.

I think the diffusion versus llm distinction is nonsense. I commented on the Cully Hill Boys post yesterday (and got downvoted today) because I'm genuinely impressed with the tech, and the fidelity and coherence that is currently achievable. Uncanny valley issues from previous model generations seem mostly gone, and the results can be "engaging." Looks like there's an actual process behind this kind of generative output, and the quality of the final product will by the way still depend on an "artistic" idea and vision, and technical execution, i.e. how much one invests in preparation (asset generation, prompting) and post-processing (editing, kill your darlings). Of course, there's also utter crap to be found on that higgs website, loads of it. There will always be low-effort slop, especially when the tech gets democratized like this, but in general I think the current sota is a big deal. It is unfortunate that it will probably cause the death of a large segment of a creative industry, if it hasn't already. We'll see how it plays out, but if my name was Boris, I'd probably say that "entertainment is largely solved."

>Every frame is generated. The actors are real and signed, but nothing was filmed with them [...] There are no sets and no camera crews. [...] The entire film was generated in Seedance — every shot, all video and speech. Faces and character sheets: Soul Cinema. Edits, reverse angles and point changes: Seedream and Nano Banana. Prompts: Claude, in two chats — one for image prompts, one for video prompts.

Color me impressed!

dupe: https://news.ycombinator.com/item?id=49448137